基于誤差擴散的指紋匹配算法model

1、中文模板基于誤差擴散的指紋匹配算法郝瑛 譚鐵牛 王蘊紅中國科學院自動化所 模式識別國家重點實驗室 北京 摘 要：匹配模塊是指紋身份鑒別系統(tǒng)中的核心模塊。在采用細節(jié)特征描述指紋的系統(tǒng)中，指紋匹配問題就轉(zhuǎn)化為點模式匹配。關(guān)鍵字：誤差擴散、指紋匹配、Hough變換1 概 述給定兩個指紋特征模板（參考模板和輸入模板），匹配過程通常給出兩個指紋模板相似程度的度量。同時，匹配還要設定一個門限，用來確定兩個模板是否是從同一個手指提取的。在點模式匹配的理想情況下2 算 法2.1 特征點的相似度度量在介紹特征的相似性度量之前，首先介紹一下指紋特征的表示。在特征提取過程中，為了去除虛假特征點，我們在細化圖像上對每

2、個檢測出的端點和分叉點所在的紋路進行跟蹤。跟蹤的結(jié)束條件有兩個：跟蹤長度達到預先設定的某個值或者遇到另外一個特征點。2.1.1 對應點估計對應點估計的目的是要找到一對或者若干對最為可靠的對應點。我們首先根據(jù)上一節(jié)定義的相似度度量，找出所有匹配上的端點對和分叉點對，然后使用Hough變換的方法找出最可靠的匹配點對作為對應點。下面將對具體的算法進行描述。圖5. 兩幅指紋圖像的匹配結(jié)果，相似度：0.739表1中列出了相同手指的不同樣本以及不同手指匹配值的均值。表1. d 以及相同手指和不同手指匹配值的均值和方差數(shù)據(jù)庫d均值（相同手指）方差（相同手指）均值（不同手指）方差（不同手指）NIST-246.

3、5688.3214.935.659.76參考文獻1 Anil K. Jain, Lin Hong, Sharath Pankanti, Ruud Bolle, “An Identity-Authentication System Using Fingerprint”, Proc. of the IEEE, Vol. 85, No.9, 1997.2 A. Ranade, A. Rosenfeld, “Point Pattern Matching by Relaxation”, Pattern Recognition, Vol. 26, No. 2, pp.269-276, 1993.3 D.

4、Skea, I. Barrodale, R. Kuwahara, R. Poeckert, “A Control Point Matching Algorithm”, Pattern Recognition, Vol. 26, No.2, pp.269-276, 1993.4 J. P. P. Starink, E. Backer, “Finding Point Correspondence Using Simulated Annealing”, Pattern Recognition, Vol. 28, No.2, pp. 231-240, 1995.5 Li Hua Zhang, WenL

5、i Xu, “Point Pattern Matching”, Chinese Journal of Computer Science, Vol. 22, No. 7, 1999.6 A. K. Hrechak, J. A. McHugh, “Automated fingerprint recognition using structural matching”, Pattern Recognition, Vol. 23, pp. 7893-904,1990.7 Xudong Jiang, Wei-Yun Yau, “Fingerprint Minutiae Matching Based on

6、 the Local and Global Structures”, Proc. of 15th ICPR, pp. 1038 1041, 2000.8 Z. Chen, C. H. Kuo, “a Topology-Based Matching Algorithm for Fingerprint Authentication”, Proc. of 25th Annual IEEE International Carnahan Conference on Security Technology, pp. 84-87, 1991.9 D. K. Isenor, S. G. Zaky, “Fing

7、erprint Identification Using Graph Matching”, Pattern Recognition, Vol. 19, pp. 111-112, 1986.10 Anil. K. Jain, Salil Prabhakar, Lin Hong, Sharath Pankanti, “Filterbank-Based Fingerprint Matching”, IEEE Trans. on Image Processing, Vol.9, No.5, pp. 846- 859, 2000.11 Chih-Jen Lee, Sheng-De Wang, “a Ga

8、bor Filter-Based Approach to Fingerprint Recognition”, IEEE Workshop on Signal Processing Systems (SiPS 99), pp.371 378, 1999.12 Anil Jain, Arun Ross, Salil Prabhakar, “Fingerprint Matching Using Minutiae and Texture Features”, Proc. of ICIP, pp. 282-285, 2001.13 Zsolt Mikls Kovcs-Vajna, “A Fingerpr

9、int Verification System Based on Triangular Matching and Dynamic Time Warping”, IEEE Tran. On PAMI, Vol. 22, No. 11, 2000.14 C. Dorai, N. K. Rathat, R. M. Bolle, “Detecting Dynamic Behavior in Compressed Fingerprint Videos: Distortion”, Proc. Computer Vision and Pattern Recognition, Vol. 2, pp. 320-

10、326, 2000.英文模板Automatic 3D Face Verification from Range DataGang Pan and Zhaohui Wu Institute of Computer System EngineeringZJU-Miaxis Joint Lab of Embeded and Biometrics TechnologyZhejiang Universisty, Hangzhou , P.R.Chinagpan, Abstract In this paper, we presented an automatic appro

11、ach for 3D face verification from range data. The method consists of range data registration and 3D face comparison. There are two steps in registration procedure. The coarse step conducts the normalization by exploiting a priori knowledge of the human face and facial features. Keywords: 3D face ver

12、ification, Range data1 IntroductionThe automatic face recognition based on 2D image processing has been actively researched in recent years, and many techniques have been presented. Although great strides have been made during the past three decades, the task of robust face recognition is still diff

13、icult. 2 3D face databaseOur experimental results use the 3D face data from “3D_RMA” database in M2VTS project5. The range data are obtained by a 3D acquisition system based on structured light. Figure 1: Two sample models in xyz form from the manual DB2.1 Face data registrationA 3D face recognition

14、 system generally makes up of two key parts: 3D data registration and comparison. The accuracy of the registration will greatly impact on the result of following comparison. 2.1.1 The coarse normalizationBefore the fine registration step, the coarse alignment is performed. We assumes that the given

15、range data are 3D facial models. Table 1: Comparison of equal error ratesBeumier5Oursautomatic DB session17.25%6.67%automatic DB session 27.75%6.67%automatic DB session 1-29.0%7.33%manual DB session 14.75%3.24%References1 G. G. Gordon, “Face Recognition Based on Depth Maps and Surface Curvature,” SPIE Proceedings, Vol.1570: Geometric Methods in Compu